Endothelial-derived relaxation factor nitric oxide (NO) was shown to mediate autoregulation and chemoregulation of cerebral blood flow. NO synethase immunoreactivity was demonstrated in the nerve plexus in the adventitia of the circle of Willis in primates. In a primate model of subarachnoid hemorrhage (SAH) adventitial disappeared on day 7 with the development vasospasm and did not return on day 14 with resolution of vasospasm, suggesting that NO loss plays a central role in the pathogenesis of cerebral vasospasm after SAH. Thus, direct replacement of NO should reverse the vasospastic effect of any NO loss. In the primate model of vasospasm, intra-arterial infusions of NO solution reversed arteriographic cerebral vasospasm, significantly increased cerebral blood flow, and decreased cerebral blood flow velocity. These findings further support a cental role of NO in the pathogenesis of cerebral vasospasm and suggest the potential of a regional NO therapy for cerebral vasospasm. We have explored the effects of the putative agents of vasospasm, oxyhemoglobin and its breakdown product methemoglobin in cell culture. These cultures studies are used to examine the possibility that vasospastic agents (e.g., endothelin) may be released from tissues exposed to oxyhemoglobin and methemoglobin. The first studies in this series have shown that astrocytes die when exposed to oxyhemoglobin in culture. This suggests that (1) oxyhemoglobin-induced endothelin release is unlikely to underlay cerebral vasospasm, and (2) cerebral injury and the production of seizures after intracerebral hemorrhage may result from distinct glial toxicity induced by oxyhemoglobin. A specific type of cranial dural arteriovenous fistulas was identified and shown to be treated effectively by simple interruption of the intrathecal venous drainage, a much simpler and safer procedure than the surgical procedure previously used to treat these patients.